patcon has quit [Read error: Connection reset by peer]
seafood has quit [Ping timeout: 245 seconds]
seafood has joined #ipfs
mguentner2 has joined #ipfs
ppham has joined #ipfs
ppham_ has joined #ipfs
ppham__ has quit [Ping timeout: 245 seconds]
mguentner has quit [Ping timeout: 264 seconds]
cemerick has quit [Ping timeout: 245 seconds]
Hoosilon has joined #ipfs
jedahan has joined #ipfs
ppham has quit [Ping timeout: 264 seconds]
jedahan has quit [Client Quit]
<whyrusleeping>
substack: do you have two ipfs nodes running? Or different ipfs paths set up? That's commonly how I end up with things hanging when I know I have them, running things against the wrong daemon
zaden has joined #ipfs
ylp has quit [Ping timeout: 260 seconds]
ylp has joined #ipfs
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
ylp has quit [Ping timeout: 260 seconds]
seafood has quit [Quit: seafood]
pfrazee has quit [Remote host closed the connection]
pfrazee has joined #ipfs
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
pfrazee has quit [Ping timeout: 252 seconds]
ylp has joined #ipfs
muvlon has quit [Ping timeout: 258 seconds]
<substack>
whyrusleeping: one daemon is running
ylp has quit [Ping timeout: 260 seconds]
muvlon has joined #ipfs
ylp has joined #ipfs
zaden has quit [Read error: Connection reset by peer]
ylp has quit [Ping timeout: 260 seconds]
ppham has joined #ipfs
ppham__ has joined #ipfs
ppham_ has quit [Ping timeout: 240 seconds]
wallacoloo____ has joined #ipfs
ppham has quit [Ping timeout: 268 seconds]
wallacoloo___ has quit [Ping timeout: 260 seconds]
ylp has joined #ipfs
muvlon has quit [Quit: Leaving]
wallacoloo___ has joined #ipfs
wallacoloo____ has quit [Ping timeout: 264 seconds]
ulrichard has joined #ipfs
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Kubuxu>
substack: update to 0.4.4, 0.4.3 includes bug that might lead to loss of some pins
infinity0 has quit [Remote host closed the connection]
<substack>
I'm using 0.4.4
infinity0 has joined #ipfs
<Kubuxu>
restart the node, run `ipfs refs -r THEHASH` and then run `ipfs bitswap wantlist`
<Kubuxu>
in separate terminal
ylp has quit [Ping timeout: 260 seconds]
woppee has joined #ipfs
<woppee>
Hey IPFS is a massive project, but I just wanted to know if there is any thumb rule regarding when devs should start breaking up structure and separate code out as npm modules
<woppee>
like say I have a project with 10, 000 loc, how to decide whether i should split things in different private npm modules and then require them?
<substack>
Kubuxu: thanks, trying now (there are a lot of refs scrolling by)
<substack>
I should expect nothing less from an archive with 200k+ files
ZaZ has joined #ipfs
muvlon has joined #ipfs
<kumavis[m]1>
null_radix, daviddias around?
<kumavis[m]1>
i have an eth-trie stumper
uu has joined #ipfs
<daviddias>
kumavis[m]1: I am, what's up?
<kumavis[m]1>
daviddias: i think i just found a heuristic that will solve my problem, but
<kumavis[m]1>
eth-trie nodes will end before the whole path is consumed
<kumavis[m]1>
and they dont bother saying what the remaining path is
<kumavis[m]1>
so a with a path like`block/state/a/b/c/d<trie ends here>/e/f/balance`
<kumavis[m]1>
when you get to the end you dont actually know how much of the remaining path (`/e/f/balance`) is supposed to be represented by that node
<haad>
good morning
<kumavis[m]1>
even though the keys for the trie are fixed length (say 20 bytes), you dont know where in the key you are when you hit the end
ylp has joined #ipfs
<kumavis[m]1>
heres my heuristic: paths stop being single characters
<kumavis[m]1>
so when you reach a tree leaf, consume the path until you hit a section where length>1
rendar has joined #ipfs
<kumavis[m]1>
well actually tx's will have { v, s, r } values for their signatures
<kumavis[m]1>
so if we limit it to valid half-bytes, it should work
wallacoloo____ has joined #ipfs
<kumavis[m]1>
previously this was not an issue bc i was treating the whole trie key paths as an non-deliminated string
wallacoloo___ has quit [Ping timeout: 246 seconds]
infinity0 has quit [Remote host closed the connection]
<kumavis[m]1>
did that make any sense?
<daviddias>
not sure if I get what means 'trie ends here'
infinity0 has joined #ipfs
<daviddias>
what happens then?
<kumavis[m]1>
we reach a leaf node
<daviddias>
you mean 'you have to jump to another node'?
<kumavis[m]1>
well, yes. bottom of the trie, we've reached our value, which is an external link to an eth-account.
<daviddias>
isn't that the same as resolving through a path that has links to other ipld formats?
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
Kingsquee has quit [Quit: Konversation terminated!]
ianopolous has joined #ipfs
infinity0 has quit [Ping timeout: 252 seconds]
john34 has joined #ipfs
PseudoNoob has joined #ipfs
greeny___ has quit [Quit: Bye.]
infinity0 has joined #ipfs
ianopolous has quit [Read error: Connection reset by peer]
<alterego>
Oh, it has to be undder src/github.com/ipfs/go-ipfs .. lame
ppham has joined #ipfs
<Kubuxu>
yeah
<Kubuxu>
yeah, due to self references to sub packages
ppham_ has quit [Ping timeout: 245 seconds]
<alterego>
urgh :P
<alterego>
symlink it is
<alterego>
Kubuxu: so, I have a pending PR on go-multiaddr-net, I have modifications in go-ipfs that depend on it, how the hell am I meant to manage this dependency properly.
<Kubuxu>
two, it won't be easy as go-multiaddr-net is used though many deps, we have tool for that WIP but easiest way will be wait for big dep update
<Kubuxu>
s/though/through
<Kubuxu>
are you creating it for the cornet PR?
<alterego>
Yes
<Kubuxu>
I don't know if cornet support for UDP will fly right now
<alterego>
I can remove UDP support, then it'll build, then add it when that PR goes through
<Kubuxu>
I think I said it in the past: we don't have stream, NAT and crypto in libp2p yet for UDP, and without that UDP won't be the thing and I don't think it is responsive to add UDP support when we don't have those feature (especially crypto).
<alterego>
UDP is pretty essential to my use case, but nvm :P
<Kubuxu>
:|
<Kubuxu>
we are working on unreliable channels, this is one of main focuses of lgierth right now
<alterego>
Yeah, and I said all my UDP requirement here is to have a UDP listening port on the corenet dialing end.
<alterego>
So it's not got anything to do with transport across the stream, it's opening a local UDP socket for applications that only work with UDP to proxy their data over the TCP peer link.
<Kubuxu>
aah
<Kubuxu>
hmm
<alterego>
That PR is a good start though for what you're talking about, it's just for me I use it to handle the case of binding a local UDP socket for app handlers.
<alterego>
(I.E. WebRTC) :P
<Kubuxu>
I see
* Kubuxu
is thinking
kenshyx has joined #ipfs
<alterego>
I'm going to do it as two separate commits anyway. And we can open the discussion in the PR
<Kubuxu>
sure
espadrine has joined #ipfs
seafood has joined #ipfs
draynium has joined #ipfs
ecloud is now known as ecloud_wfh
<alterego>
Kubuxu: how do I run go-ipfs tests locally?
<_mak>
when I download a file from the ipfs gateway, is there any way to generate a multihash of that file to check if the file multihash matches the filename (ie: integrity is ok)?
<Kubuxu>
TEST_NO_FUSE=1 GOTAGS=-nofuse make test_short
<alterego>
Thanks
<Kubuxu>
alterego: gotags without the -
<alterego>
k
<Kubuxu>
you don't need it after all
<Kubuxu>
we changed that
<Kubuxu>
just TEST_NO_FUSE
<Kubuxu>
_mak: you can try `ipfs add -n THEFILE`
<Kubuxu>
but it is not guaranteed to give you the same hash
<Kubuxu>
best way is to run ipfs node and download using it
<_mak>
Kubuxu: I wonder if there is any bash tool that will generate the multihash of a file
<Kubuxu>
it isn't that easy because there are many ways to chunk a file
<_mak>
the use case is in a machine without ipfs
<_mak>
hmm
<Kubuxu>
and this will influence the resulting file
<Kubuxu>
if not chunk, there are many ways to link the chunks so they create file
<_mak>
the hash I download on the gateway is the root hash right?
<Kubuxu>
yes
<_mak>
and that hash is unique to how ipfs generate the multihash
<ansuz>
daviddias: how different is brazilian portuguese from what you're used to?
<victorbjelkholm>
daviddias: hah, similar
<daviddias>
ansuz: I'm used to both
<daviddias>
we get exposed to both a lot here
<daviddias>
The same is not true if you grow up in Brazil
<victorbjelkholm>
had some issues with the graph when I was working on it, had issues with streams closing prematurely and also disconnects not being always sent
<daviddias>
victorbjelkholm: those things might be better now, when was your last npm update deps and npm fresh install?
<ansuz>
daviddias: cryptpad has pt_br support now, so I guess that works well for .pt folks then :)
<daviddias>
yeah, that is what most products/companies default too
<victorbjelkholm>
daviddias: hm, was last week sometime
Encrypt has quit [Quit: Quit]
<daviddias>
although it makes me a bit sad not seeing the pt_PT, there are more countries that speak pt_PT
<daviddias>
but to be honest, I do feel that pt_BR is way more beautiful than the original pt_PT :D
<daviddias>
victorbjelkholm: try again today :)
<victorbjelkholm>
daviddias: All right, I'll try!
<ansuz>
daviddias: we just lucked out and had someone offer to translate it, not at all a conscious choice
<daviddias>
ansuz: oh I was assuming that, no problem at all :) I was just pointing out that, even though for the people that live in Portugal it is fine, it won't be enough for the other pt_PT speaking countries :)
<ansuz>
I learn a lot about the world because of software
<ansuz>
:)
bastianilso_____ has joined #ipfs
cemerick has joined #ipfs
<daviddias>
:)
bastianilso_____ has quit [Quit: bastianilso_____]
<jbenet>
victorbjelkholm daviddias: random idea -- what if the nodes in the libp2p graph are flags that show the country of the ip addr
ivo_ has joined #ipfs
Boomerang has joined #ipfs
ppham_ has joined #ipfs
ppham__ has joined #ipfs
infinity0 has quit [Ping timeout: 240 seconds]
taaem has quit [Read error: Connection reset by peer]
<alterego>
Kubuxu: figured out why those corenet tests didn't work
ppham has quit [Ping timeout: 248 seconds]
taaem has joined #ipfs
<alterego>
Kubuxu: 100K of data is too big to send in one chunk it seems.
infinity0 has joined #ipfs
ppham_ has quit [Ping timeout: 258 seconds]
taaem has quit [Remote host closed the connection]
taaem has joined #ipfs
taaem has quit [Remote host closed the connection]
taaem has joined #ipfs
taaem has quit [Remote host closed the connection]
taaem has joined #ipfs
<jbenet>
alterego it shouldn't be. i think the block limit is 8M
PseudoNoob has quit [Remote host closed the connection]
<Kubuxu>
alterego: just got home from Uni, looking at it
<Kubuxu>
not really "just got home", I was coding with all comms disabled :p
<alterego>
:D
<alterego>
Fair enough, I think I'm going to need some help getting rid of these CI errors ..
<Kubuxu>
yeah, just looking at it
<Kubuxu>
alterego: first test is failing because you don't have description for the all option in close method
<Kubuxu>
it shouldn't really, you just exposed a bug :p
<alterego>
Oh, whoops.#
<Kubuxu>
I will fix it, but just for now add short help text to the options.
<alterego>
Yeah, that kinda was my bad really.
john34_ has joined #ipfs
<alterego>
Kubuxu: the really weird one is the sharnness test.
<_mak>
am I correct to say that ipfs is not optional to store big files with binary content?
<alterego>
_mak: that's kinda what it's designed for.
<lgierth>
Kubuxu alterego: i'm absolutely fine adding udp support to multiaddr-net without any of the rest (swarm, crypto, etc.)
john34 has quit [Ping timeout: 260 seconds]
<alterego>
Kubuxu: that's great :) Check that PR then :D
<alterego>
Erm lgierth ^ :P
<lgierth>
:)
<_mak>
alterego: I thought it was optimal for files with text content that can be deduplicated easily
<alterego>
_mak: there's no real differentiation with IPFS.
<r0kk3rz>
_mak: binaries can be deduped as well
<Kubuxu>
I will look at it, finishing the option-lacking-desc patch
<_mak>
I ask because I have ~400GB of binary files pinned and any operation I do will take many hours, I've noticed that the quantity of blocks is ridiculously high compared to how many files I have
<_mak>
Yday night I ran a gc and it is still running after 12+ hours (on a quite good machine)
<Kubuxu>
we are working on it, some of our repos are in TBs so we know that
<_mak>
ipfs pin ls will also take many hours
<Kubuxu>
test out the master right now or one of pre-releases, there were some improvements in this regard
<_mak>
r0kk3rz: how can a binary file be deduplicated? if you split files in chunks of 256kb isn't this the same as having a 256kb hash?
<alterego>
_mak: files are split into many blocks when they get large. this is for blockswapping between peers like bittorrent.
<lgierth>
yeah 256KB blocks
<_mak>
Kubuxu: I see, good to know, I thought I was making a bad use of it, I'll test the pre-release
<r0kk3rz>
_mak: block level dedup
<_mak>
what are the chances of having 2 256kb blocks of binary content with the exact same content?
<alterego>
Kubuxu: I guess that empty help text is probably what broke the sharnness test too.
<alterego>
Kubuxu: hrm, or not, failed again.
<Kubuxu>
alterego: there are few mistakes in sharness that are hiding problems, I will comment in the PR
<alterego>
Okay, thanks.
kode54 has quit [Ping timeout: 252 seconds]
<daviddias>
victorbjelkholm: oh man! look at those graphs coming out from jsipfs tests <3 <3 <3
kode54 has joined #ipfs
Encrypt has joined #ipfs
<alterego>
Kubuxu: thanks for the input, I'll work through them in a bit.
<Kubuxu>
alterego: if you have, try using bsd-nc when testing those tests, this might also be part of the issue.
<tiago>
no debian package yet?
<alterego>
Kubuxu: there is no other test that uses nc in listening mode.
<tiago>
or am I missing something? ;)
<alterego>
It has to be &
boggle has joined #ipfs
<Kubuxu>
I know it has to be &, what I mean is: the ipfsi and the nc line below it
<alterego>
Oh, right.
<alterego>
To break if the command fails right?
<alterego>
Okay :)
infinity0 has quit [Remote host closed the connection]
<Kubuxu>
yeah
anonymuse has joined #ipfs
infinity0 has joined #ipfs
<lgierth>
Kubuxu: can you cover for whyrusleeping in leading the go-ipfs call? i can otherwise
<Kubuxu>
yeah, I can do it.
<lgierth>
cool++
infinity0 has quit [Remote host closed the connection]
<Mateon1>
_mak: Not that low, when you add multiple versions of the same firmware (e.g. for different devices) there is a surprising amount of deduplication even with this braindead chunker. One of the guys on the team added a bunch of versions of OpenWRT, most had deduplicated 0x00 blocks, but even without 0x00 blocks they had decent deduplication.
<Kubuxu>
or stuff like minified js, I had reduction to about 80% of source, not incredible but good
infinity0 has joined #ipfs
zombu2 has quit [Ping timeout: 256 seconds]
<lgierth>
we'll see some great dedup with npm-on-ipfs too
<lgierth>
the issue with smarter chunkers like rabin fingerprinting (ipfs add -s rabin) is that they're less efficient about disk blocks
john34 has joined #ipfs
zombu2 has joined #ipfs
john34_ has quit [Ping timeout: 260 seconds]
ZaZ has quit [Read error: Connection reset by peer]
boggle has quit [Quit: Leaving]
<alterego>
Kubuxu: modified test now. Got 4 failures still :(
<Kubuxu>
alterego: && wati $NC_PID, but the problem is the `-p` flag, it isn't compatible with `-l`.
<Kubuxu>
just don't use it
anonymuse has quit [Read error: Connection reset by peer]
anonymuse has joined #ipfs
<Kubuxu>
just don't use `-p` it isn't needed
ppham_ has joined #ipfs
ppham__ has quit [Ping timeout: 265 seconds]
ashark has joined #ipfs
pfrazee has joined #ipfs
[0__0] has quit [Remote host closed the connection]
[0__0] has joined #ipfs
trn has joined #ipfs
bastianilso_____ has quit [Remote host closed the connection]
bastianilso_____ has joined #ipfs
JonnyNomad has left #ipfs ["leaving"]
<alterego>
Kubuxu: ah, good. Okay.
<alterego>
I'll stick the IP address in it's place as well.
<alterego>
Kubuxu: I'm guessing your comment was about peerstore import?
<Kubuxu>
imports in general, see that there are only 3 sections in that doc
<Kubuxu>
stdlib, internal and external
<alterego>
Yup
<alterego>
Okay, reordered and made pstore peerstore
leeola has joined #ipfs
gully-foyle has quit [Ping timeout: 258 seconds]
zanadar has joined #ipfs
domanic has joined #ipfs
zanadar has quit [Client Quit]
shizy has joined #ipfs
gully-foyle has joined #ipfs
zanadar has joined #ipfs
infinity0 has quit [Ping timeout: 246 seconds]
zanadar has quit [Client Quit]
infinity0 has joined #ipfs
zanadar has joined #ipfs
zanadar has quit [Remote host closed the connection]
clavi has quit [Ping timeout: 248 seconds]
clavi has joined #ipfs
ulrichard has quit [Remote host closed the connection]
PseudoNoob has joined #ipfs
Nukien has quit [Remote host closed the connection]
Nukien has joined #ipfs
r04r has quit [Ping timeout: 248 seconds]
<alterego>
Kubuxu: So what about these codecov tests?
r04r has joined #ipfs
<Kubuxu>
they don't include sharness yet
<Kubuxu>
don't worry about them
ylp has quit [Quit: Leaving.]
<Akaibu>
whyrusleeping: do you have any plans for an IPFS OS? Will it be somewhat like Plan 9?
<alterego>
Teehee
<alterego>
Kubuxu: okay, well I made all those changes then :)
john34 has quit [Ping timeout: 260 seconds]
<Kubuxu>
it works, I and few others have to review it and talk about it a bit more, expect some traffic in the PR :p
<richardlitt>
sprint-helper: next
<sprint-helper>
The next event is "All Hands Call", in 26 minutes.
<alterego>
lgierth: is there any chance I can get the UDP thing in with my UDP dependant corenet code in the next rel?
<Kubuxu>
alterego: hard, it is completely new feature (corenet), we are already over a month after our planned review but to brighten your day, in 2017 we plan releases every 2 weeks https://github.com/ipfs/notes/issues/203
infinity0 has quit [Ping timeout: 246 seconds]
<alterego>
Kubuxu: isn't that why it's under experimental? :P
<alterego>
tbh, I'm happy if it is just in master, soonish :D
lkcl has quit [Ping timeout: 246 seconds]
<Kubuxu>
it is, but we are really already into release process, IDK if we will have time to even review it before end of the year, sorry
infinity0 has quit [Remote host closed the connection]
<jbenet>
lgierth: maybe file an issue saying something like: "add a way to turn the animation off (make network static), that uses a cookie to remember my preference" or something
<alterego>
lonely, so lonely .. I have nobody, to be my p2p peeeeers. :)
<jbenet>
alterego: make sure you do the "allow unsafe scripts" thing
<alterego>
I'm using ff
<alterego>
I'll try in chrome
<lgierth>
oh yeah... https and HSTS
<lgierth>
and a lack of /tls :)
<alterego>
I SEE PEERS! :D
<alterego>
FRIENDS
<alterego>
That is epic
infinity0 has joined #ipfs
<victorbjelkholm>
daviddias: you removed the directory explorer from ipdendro? :(
gavinmcdermott has joined #ipfs
gavinmcd_ has joined #ipfs
gavinmcdermott has quit [Read error: Connection reset by peer]
tsnieman[m] has joined #ipfs
se3000 has joined #ipfs
<lgierth>
what's ipdendro
<victorbjelkholm>
did hackmd.io just die?
<victorbjelkholm>
doesn't work for me...
<alterego>
Someone tell him about my STUN by pass through corenet :P
<jbenet>
it's ok to be polite :) -- but yes, feel free to just ask
old_kanye has quit [Ping timeout: 260 seconds]
<pjz>
victorbjelkholm: that... should work...
<old_kanye_>
Someone suggested that they want an operating system based on ipfs
<victorbjelkholm>
pjz: guessing it shouldn't have the production.
infinity0 has joined #ipfs
<victorbjelkholm>
in the beginning, and probably you have a host setup for that hostname
<pjz>
victorbjelkholm: ah, hm, yah, I'll have to look into that! It was working yestrday :)
<old_kanye_>
But my question is, isn't that pointless since you can just mount a filesystem? You don't need an operating system for it right?
<alterego>
You need an OS to mount something.
<alterego>
Otherwise you've just got a box with pretty lights.
<ansuz>
:d
<kevina>
Is go-ipfs still on?
<daviddias>
victorbjelkholm: I didn't
<Kubuxu>
kevina: we are just starting
<daviddias>
It is still there
<daviddias>
Need to converge the PR's
<old_kanye_>
Someone on that thread said "Basically the idea is to remove the concept of local vs remote files at the OS level. So as a user of the OS the file browser and the web browser are the same application. They shouldn't even be able to tell if something is on their computer or not. This could be implemented on HTTP but only as a prototype."
<old_kanye_>
Is this correct?
<daviddias>
I did that thing that same night, before the morning actually
<daviddias>
next time I'll demo with the explorer :)
<victorbjelkholm>
daviddias: ah, I see! No worries
<jbenet>
old_kanye_ yes, that's right. i've been wanting this since the inception of IPFS. it will take a bit of time to get there with high perf-- but yes.
<jbenet>
you would also get a hash-chain of the state of the entire OS FS, which is pretty cool. can re-start at a verified hash every time you reboot, and avoid injected malware
<flyingzumwalt_>
does anyone know how to run the sprint-helper?
<flyingzumwalt_>
to post the info for go-ipfs call?
todder has quit [Ping timeout: 245 seconds]
<pjz>
jbenet: Gobo has a filesystem based on content hashes
<flyingzumwalt_>
Watch Stream: TBA soon by dignifiedquire
<victorbjelkholm>
my bad, notes url is wrong as well, can't find the existing one
<victorbjelkholm>
lgierth: Kubuxu dignifiedquire what hackmd are you on?
<frood>
jbenet: starting from a known hash would essentially create an amnesiac FS (unless you explicitly update), right?
<victorbjelkholm>
guessing there is no notes being written...
<jbenet>
frood: yeah that's fine for many applications
<old_kanye_>
jbenet: Thank you :), but once again, sorry but i'm still having trouble understanding how an IPFS OS will be different from the oses we have now. Since IPFS is a filesystem, can't you just use a normal OS? Why is it a prototype on a normal OS?
<jbenet>
frood: invert the "/ and /tmp" relationship.
<jbenet>
frood: and certainly for servers/hypervisors ...
<jbenet>
old_kanye_ im not suggesting it has to be a new os, but i _would_ like to get rid of a lot of cruft some day.
<frood>
known-good states are known to be good. :)
<jbenet>
old_kanye_ a ton of assumptions about how the fs is structured, or that most of it is mutable
<old_kanye_>
Ohh that's why it's a prototype on the oses we have now, The filesystem on an ipfs os will be immutable while the ones we have now are mutable... So when i mount ipfs on my linux with fuse, my os just assumes it's mutable?
<alterego>
old_kanye_: no, it's readonly iirc
<alterego>
old_kanye_: because you can't just write a file when you haven't derived the name from the hash
<alterego>
old_kanye_: likewise when you "edit" a file, the hash changes.
<dignifiedquire>
david, sadly no :/ go-ipfs sends garbage bytes over the wire through json (http api)
<dignifiedquire>
it's literally not decodable anymore
<dignifiedquire>
it's not about understanding protobufs from go, it's about what go does to the messages when sending them through the http-api
<daviddias>
ah, roger
gavinmcd_ has quit [Read error: Connection reset by peer]
<old_kanye_>
alterego: Sorry I just heard about ipfs a couple of days ago, I don't know what "you can't just write a file when you haven't derived the name from the hash" means, I'm guessing an IPFS OS should be able to do that...
<daviddias>
though it was the decoding of the protobuf
<flyingzumwalt_>
FYI go-ipfs discussion is running long, covering some important points
<daviddias>
because string and bytes are different things
<Kubuxu>
what go-ipfs is sending there?
<Kubuxu>
binary multihash?
<daviddias>
binary cid
<Kubuxu>
then whyrusleeping was at least partially wrong ;p
infinity0 has quit [Remote host closed the connection]
<Kubuxu>
like on protobuf lever those are the same tags, in Go it doesn't matter much (as string is fully binary) but the protobuf recommendation is due to languages no handling binary strings
infinity0 has joined #ipfs
se3000 has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<keks>
richardlitt: the streaming link in the sprint issue still links to that issue
<Kubuxu>
you get the shortest hash fragment that isn't yet used
<Kubuxu>
then reference it from domain `fs:/ipns/short.eth/abc` in future it might be possible
<kevina>
BTW where is whyrusleeping today?
infinity0 has joined #ipfs
<herzmeister>
yeah, i've been long in the blockchain space, and let's say i'm not betting on ethereum, but yeah, a short identifier system would require some sort of blockchain for duplicate checking, but first and foremost i also don't want links to break because of user-chosen names
ligi has quit [Ping timeout: 252 seconds]
<lgierth>
kevina: chistmas and family
ulrichard has joined #ipfs
<kevina>
lgierth: so he won't be back to after the holidays?
<jbenet>
he'll likely be around some
ppham has quit [Remote host closed the connection]
<jbenet>
but dont expect major things during holiday season -- people need rest :)
<lgierth>
crunch week for me personally
<lgierth>
i have like a dozen 90%-finished things
<kevina>
:)
<old_kanye_>
jbenet: Thanks a lot for your time :), I'll have to think more about this ipfs os. It's okay if i come back and ask more questions, right?
<flyingzumwalt_>
If you think more people should be reading this, look for it on HackerNews and upvote https://news.ycombinator.com/newest The title is "Instructions for Saving Endangered Data: It’s Time to Get Decentralized "
<victorbjelkholm>
I actually mention ipld.io, libp2p.io and multiformats.io there
<lgierth>
victorbjelkholm: yep!
infinity0 has joined #ipfs
<gavinmcdermott>
dignifiedquire - I'll dive into a review here this afternoon when I get some space! Thanks for hopping in; especially with the shift to the event-emitter interface - the object stream was a small pain last week
<victorbjelkholm>
gavinmcdermott: what David mentioned, kubernetes-ipfs: https://github.com/victorbjelkholm/kubernetes-ipfs currently just something I hacked together over the weekend, I can give you a demo after the libp2p call if you want
<gavinmcdermott>
victorbjelkholm - absolutely!
<victorbjelkholm>
I like it that jbenet appeared in the libp2p call 2 seconds before it got stopped :D\